The Platform as a Service Model for Networking
|
|
- Clyde Phelps
- 5 years ago
- Views:
Transcription
1 The Platform as a Service Model for Networking Eric Keller Princeton University Jennifer Rexford Princeton University Abstract Decoupling infrastructure management from service management can lead to innovation, new business models, and a reduction in the complexity of running services. It is happening in the world of computing, and is poised to happen in networking. While many have considered this in the context of network virtualization, they all focus on one model overlaying a virtual network of multiple virtual routers on top of a shared physical infrastructure, each completely isolated from the others through the use of virtualization. In this paper we argue for a different approach, where those running the service are presented with the abstraction of a single router in order to enable them to focus solely on their service rather than worrying about managing a virtual network as well. We discuss the abstraction of a single router, and the challenges of mapping the collection of abstract routers (from different parties) to the distributed and shared physical infrastructure. 1 Introduction In the world of computing, a shift has begun towards the use of infrastructures which are hosted and shared (i.e., cloud computing). This has increased the level of innovation by enabling companies to come out with new web services for less cost, created new business models where a party can lease out slices of servers on demand with a pay-per-use model, and even simplified management in private (non-hosted) networks by enabling a company to more easily run independent services on its own servers. We believe that the same will be true of networking where many applications would greatly benefit from innetwork functionality beyond the basic connectivity offering of today. This is already underway, as many researchers have been exploring this model for years new business models decoupling who owns the physical infrastructure (infrastructure provider) from who runs and configures the service (service provider) [1, 2, 3], shared testbeds using virtualization to enable innovation in networking research [4, 5], and capitalizing on virtualization to ease the burden of running multiple services in a single ISP [6, 7]. However, these studies have focused only on a single model that of a network of virtual routers, where one can lease or use a slice of physical routers connected together via partitioned links. Taking from the terminology used in cloud computing, we characterize this as analogous to the infrastructure as a service (IaaS) model since it simply slices the physical resources. However, in the virtual networks model, the virtual network needs to be managed the same way the physical network is e.g., dealing with failure and link congestion. In this paper we explore an alternative model the platform as a service (PaaS) model. Instead of exposing the underlying physical network and topology, we propose the abstraction presented to customers should be a single router. As routers today include both route processing blades as well as general compute blades, a trend that will likely continue, the single router abstraction is general enough to cover all in-network functionality (i.e., not just routing). This model benefits both the customers, as they can run their service without needing to manage a physical network, and the infrastructure providers, as they can use their platform as a source of differentiation. In the remainder of the paper we first discuss how applications would benefit from more in-network functionality than what is offered by ISPs today (Section 2). We then discuss why virtual networks place too much burden on the customer (Section 3). Next, our proposed router platform model is presented where we discuss both the abstraction and the challenges of realizing it on a distributed and shared network (Section 4). We follow that with a discussion on how the router platform model can help ease private network management (Section 5) and conclude (Section 6).
2 2 Basic Connectivity is Not Enough What ISPs offer today is basic connectivity i.e., the ability to communicate with any other IP address at a given bandwidth. However, as many have argued, this is not enough as many applications can benefit from innetwork functionality. Here, we briefly discuss a few of these (this list is by no means exhaustive). Customer controlled routing: Despite the large path diversity available in ISPs, the mechanisms of the Border Gateway Protocol (BGP), used for inter-domain routing, are such that a single route must be chosen at each router and each route reflector. This means that this diversity is not made available to customers, instead forcing the ISP to announce the same route to each customer. At the same time, each customer is different and has its own preferences of what constitutes a route as being the best. One may desire a low latency path, another may desire a path that does not go through certain countries, while a third may desire a low cost path. If the ISP went beyond basic connectivity, they could in fact enable different neighbors to receive different routes by giving the customer control over route selection (either as coarsegrained selection, such as prefer lowest latency, or by giving complete control to the customer) [8] [9]. Cloud computing: Cloud computing is an exploding business where enterprises and providers of web services lease compute resources from a cloud provider. However, the customer is simply given servers and the capability to reach each server (both externally and internally). The customer does not have any ability to configure the network components (routers and switches), as they would in their own private facilities [10]. This limits the middle-box type functionality (such as a firewall or intrusion prevention system) that enterprise networks commonly have. It also hinders the ability to use virtual private networks (VPNs), which enterprises commonly use for inter-site communication, as existing cloud providers do not provide enough capability to enable the customers to securely attach to a VPN endpoint [11]. Gaming and Live video streaming: Online games, whether action oriented first-person shooters or more exploratory virtual worlds, have a great need for the ability to efficiently distribute updates to all of the players. Live video streaming also requires an efficient distribution of the stream. Publish/subscribe and multicast are ideal for both of these applications. These applications also share the need for on-path processing (e.g., to aggregate game updates or to transcode the video stream). However, an efficient distribution or on-path processing are not part of the basic connectivity offering. Network monitoring: Many applications have the need to monitor the condition of the network (e.g., the availability, congestion, or latency of paths) in order to troubleshoot or perform, for example, server selection. This includes a number of techniques such as observing BGP updates or performing ping tests. Given the distributed nature of the Internet, monitoring servers must be setup throughout the world. For this reason, in-network visibility is a more accurate and efficient way to monitor the condition of the network than those limited by the basic connectivity offering. 3 Virtual Networks Are Too Much To go beyond basic connectivity, ISPs can add functionality to their network and offer it to their customers. However, they are limited to the geographic foot print of their network and innovation is limited to a single party (the ISP). Overlay networks have arisen as a way around this, enabling 3rd parties to run specialized network functionality (e.g., content delivery networks) by setting up servers throughout the world and using the Internet as links between the servers. Still, each of these are one off deployments and therefore researchers have proposed virtual networks as a generalization which enables multiple parties to create custom networks on a shared physical network substrate [1, 2, 3, 4, 5]. In the virtual networks model, multiple virtual networks can co-exist on a shared physical substrate through the use of virtualization. A virtual network is a collection of virtual routers connected in a topology via virtual links. A virtual router is a slice of a physical router (i.e., the physical router s resources are partitioned and each virtual router is isolated from one another). Virtual links are created through partitioning the physical link s bandwidth and creating tunnels when adjacent virtual routers are not located on adjacent physical routers. Those wanting to create a custom network (either for themselves or to sell as a service to others) can do so by leasing resources from an infrastructure provider (the party that owns and maintains the physical network). However, while overcoming the limitations of basic connectivity, there are problems with the virtual networks model. These problems are centered around the fact that managing a virtual network is similar to managing a physical network. There is limited bandwidth on the virtual links between virtual routers, and so the customer must engage in traffic engineering to optimally use the virtual link bandwidth. Further, as there is a one to one mapping of virtual routers to physical routers, the failure of a physical component (line card, route processor blade, or link) is visible to the virtual networks that use those components. Because of this, the customer must be able to cope with failure (e.g., by providing redundancy). Making each customer manage the virtual 2
3 network, with the same burden as managing a physical network, is unappealing to the customers, who really only want to run their service. At the same time the virtual network model may not be desirable for the infrastructure provider either. First, when the same party that runs the services also owns the physical infrastructure, planned maintenance events can be coordinated (e.g., reconfigure the service to route around a router that is going to be shut down). This coordination is made more difficult when the two parties are different. Virtual router migration [7] is a possible way to deal with this, yet it still takes time and depending on what the customer s router is doing, may not be applicable. Second, partitioning is an inefficient use of resources. The infrastructure provider has no control over, for example, how much traffic each virtual router will send to each other virtual router, and therefore must allocate enough resources so the customers receive the service level they are promised. Third, and perhaps most important, the infrastructure provider is simply offering a commodity service there is little opportunity to differentiate themselves from other infrastructure providers other than cost. This is unappealing to the infrastructure providers, and something the ISPs are grappling with in today s basic connectivity offering. 4 A Router Platform is Just Right Virtual networks go too far in compensating for the limitations of the basic connectivity model. Instead, we propose presenting customers with a platform that is decoupled from the physical infrastructure making it so the customers only have to manage their services and the infrastructure provider has freedom to manage the physical infrastructure. This platform will look very much like a router today, making it both familiar to network operators who are used to specifying routing policies in terms of router configuration, and generic enough to cover any type of in-network functionality. In the remainder of this section we discuss this single router abstraction and how it is realized on a distributed and shared physical infrastructure. 4.1 The Single Router Abstraction What the customer really cares about is being able to configure its specific policies and how packets are handled. As such, we propose that they are provided the view of a single router. Important to note is that we view a router in a broader sense than what they are traditionally thought of. That is, routers are becoming more and more powerful and converging towards a large, heterogeneous distributed system (today s high end routers are already there). They have many line cards to perform Figure 1: The single router abstraction. data plane functions, route processing blades which runs routing software to calculate paths through the network, general purpose processing blades allowing general software to be run (control or packet handling), and a switching fabric that connects all of the cards together. From this, the platform we propose as the abstraction presented to the customer is as show in Figure 1, covering the three main processing components in a router the route processing, the data plane, and the general purpose processing. Interactive program : Rather than a static configuration file, we envision the programming environment being an application programming interface (API) in which case the customer would provide an executable program (e.g., a Python script). This program would include an initialization routine, which would use each component s API to configure the router in its initial state. Then, the program can dynamically update this configuration. This may come as a result of receiving a message (e.g., from one of the customer s servers) indicating a change to be made. Or, it may come as a result of a network event the definition of which is broad and will be specific to each component, but may include things like links going down or cost of using a link changing. To handle these events, the customer s program will provide callback functions and register itself as being interested in a particular event. Routing : The routing component provides the ability to customize path selection. Much as with today s configuration mechanism of routers, customers will be able to specify sessions with neighboring routers. This will include both sessions to the customer s own routers (e.g., in its own facilities, or run by another infrastructure provider) as well as sessions to routers that the infrastructure provider has a business relationship with (either its upstream provider or a peer). Note that, specifying a session is for configuration only sessions are shared among all customers. With this configuration, the customer can also specify policy which determines which routes are preferred, what modifications to make to routes, and which neighboring routers to announce up- 3
4 dates to 1. To be able to select routes, the customer needs to know information about the various links it has the option to choose among. First, it needs to be able to query what the available links are. Second, it needs to know certain metrics for each link such as relative cost. Previously, the ISP would choose a given link (e.g., the cheapest) and that is what the customer would use. With this, they have the option of using a more expensive link if they desire. We envision that these metrics can and will change, and this can be notified to the user through the event mechanism, in which case the user can update the configuration. Data plane : The configuration of today s routers includes configuring the data plane functionality in addition to the routing policies. We view them as separate functions and as such treat the data plane as a separate module from the routing. For this, we are not suggesting that the customer s program will interact directly with the line cards. Only that a sub-set of the capabilities of a line card will be exposed to the customer s program. This includes such functionality as being able to configure multicast groups or setting up access control lists (ACLs). As an extension to the current static configuration, the data plane can also have events that a customer s script can act upon. Events here are intended to be notifications about aspects the customer cares about (e.g., the traffic rate on a given link has reached a certain threshold) and not those that are handled by the routing software already (such as a link failing), though even those could be informative. General-purpose processing : The general purpose processing module is the most generic, and therefore, difficult to generalize. However, it will be presented to the user as though they are configuring a special purpose machine such as a server which performs measurement, or a firewall, or file caching storage layer, or any other functionality that benefits from being inside of the network. This is really a way for the infrastructure provider to differentiate itself from other infrastructure providers. Longer term, we believe that customers may be able to write their own software modules including per-packet handling, per-flow handling, or control only processing. This would be facilitated by virtualization technologies which would enable isolation of the module from other modules, ensuring other customers are not affected by faults in the customer s software. 4.2 The Physical Infrastructure Reality Our proposal is to provide each customer the ability to program a single router. There are two main challenges 1 For multiple links to the same neighboring AS, when all else is equal, the customer can say it does not care and the infrastructure provider is free to choose (e.g., based on location). of this when realizing it on the actual physical infrastructure: (i) that the infrastructure is distributed and (ii) that the infrastructure is shared. Each of these is discussed below Distributed Router The abstraction of a single router does not match the architecture of networks today consisting of many routers, which collectively handle the routing and forwarding workload, and a variety of special purpose machines. In fact, routers are becoming more like single box versions of the networks they run in. As such, we will generalize the definition of a physical router (as opposed to an abstract router) to be the combination of route processing blades, packet processing blades, and general-purpose processing blades, all connected via a large switching fabric. The switching fabric can either be a network of routers (e.g., running OSPF), a network of Ethernet switches, or a switch internal to the router (e.g., a Clos or crossbar). This idea of distributing the workload among many routers and special purpose machines is not really anything new, operators readily do this today. However, due to the decoupling of who owns and maintains the physical infrastructure from who configures the router, the process must be highly automated and must be able to change. The need to be able to dynamically change comes from the fact that the infrastructure provider does not know the customer s exact workload or exactly which functionality will be used. It also simplifies workload distribution since if the placement turns out to be a poor selection, it can be changed. Automatic distribution then involves four steps, each of which is briefly discussed below. Choose a placement : There are a number of factors that go into deciding where to place a certain resource. First, there is the functionality requirement of the customer s router if it is using some capability (e.g., multicast or a firewall), that processing must be done on a router that supports that capability. Second, the processing requirements of the individual parts of a customer s router must be met we can partition routing sessions across multiple route processors and we can replicate middle-box functionality in order to scale. The rate at which routing updates (for routing software) or data packets (for general-purpose software) arrive along with the estimate of how much processing each request requires, will determine how much each processor can support. Third, basic performance factors such as latency and maximum throughput need to remain reasonable. Fourth, the traffic on the internal network must be taken into account. This may come from internal control communication (e.g., ibgp traffic), data traffic going from ingress to egress 4
5 router, and data traffic to each middle-box in succession. Configure inter-processor communication : When a placement is decided upon, the inter-processor communication needs to be set up. Each middle-box type module will have it s own inter-processor communication (e.g., a caching solution might exchange what is currently being cached with one another). For routing, the inter-processor communication is an exchange of routes and is typically set up via the configuration language (e.g., specify all of the ibgp routing sessions) since the customer s view is that of a single router, it does not need to configure the ibgp sessions, they are automatically configured by the infrastructure provider. Once configured, each processor typically selects a best route locally from the sessions terminated at that processor and sends those to each other so that each will select an overall best route 2. Tune the switch : The impact on the internal network is taken into account when deciding placement. However, since it it not the only factor, there is also the need to optimize the internal network through traffic engineering. This may, for example, include adjusting IGP routing protocol parameters, such as link weights. Dynamically adapt : The placement process uses estimates to determine the best placement. This limits the effectiveness as there is limited information when performing the placement and the demands may change over time. Because of this, we view the process as a more dynamic one where the placement can change. This requires monitoring all of the key inputs (CPU usage, update frequency, traffic patterns, etc.). From this, we can re-run the placement algorithms. However, changing placement is not always easy. State may need to be transferred and the actual process may require some processing e.g., migrating a BGP session involves exchanging the routing information and then re-running routing processes [12]. Therefore, some notion of the cost of migrating a particular workload is taken into account Shared Infrastructure Going beyond supporting the single router abstraction on a distributed infrastructure, the model must also support multiple customers, each specifying their own abstract router. For this, the individual components of the distributed network need to themselves have support for supporting multiple customers. The need to support multiple customers has been studied in the context of virtual networks. However, since in our model, each customer is only allowed to configure the router, as opposed to run their own routing processes, 2 Due to BGP not being totally ordered, multiple routes may have to be exchanged. it is more efficient to simply run a single control process handling all customers routing sessions. In fact, the abstract routers cannot be completely isolated as they can share a session with a neighboring router if multiple customers specify the same neighboring router as a peer, only a single session is established (as that is what the neighboring router expects). To handle this, when a route is received, the routing software will replicate the message, tag each with a customer identifier, and process each using the particular customers policy. For route announcements that are generated as a result, the customer identifier is used to send it only over the customer s sessions. When one of the outgoing sessions is shared, the infrastructure provider will choose among the routes each customer is sending most likely, a given prefix will be announce only by one of the customers (e.g., their own prefixes). The case where sessions are unique is simply a special case of the above, where a received route does not get replicated and a sent route does not have any alternatives for the infrastructure provider to choose from. This choice to run a single routing process then impacts the data plane and general-purpose processing blades. In virtual networks, network namespaces is a mechanism to completely separate each configurable aspect of a shared packet handling process (whether in hardware or software) such that each control plane can configure their view of the data plane without any contention [13] 3. It does this through the use of contexts which essentially use the identifier of the control process as an index into the various data structures that are used in processing a packet (e.g., the forwarding table). With our proposed router platform, there is one key difference the context is exposed to the control plane software (routing software or the software presenting configuration options to the customer s program), rather than being enforced with virtualization. The implication of this is that the control software is not tied to the strict notion that there are N completely isolated routers, and N completely isolated forwarding tables, which can lead to optimizations. 5 Simplified Management of Private Networks A side effect of the router platform model is that it simplifies management of private networks. Automate configuration : As discussed in Section 4, we envision that the decoupling of who owns and maintains the equipment from who is running and configuring services on the equipment will require a great deal 3 NetNS is a Linux specific implementation, but the concept has general applicability. 5
6 of automation involving the infrastructure provider to handle every change each customer makes is too time consuming. The network operator of a private network will benefit from this, as they too can specify using the abstract single router and leave the management of the physical infrastructure to automated tools. Additionally, the operator will be able to dynamically modify their configuration, which will have an effect on the underlying physical infrastructure. Manage separate services independently : Additionally, the support in the infrastructure for multiple concurrent services to be run and configured independently simplifies the management of the network as a whole. Rather than having to run a new service on a test network and then gradually integrate it into the production network, each service can remain independent and co-exist on the production network without affecting one another. Further, different business units within the same company can manage their service independent of one another. The private cloud analogy : This simplified management has an analogy in the field of cloud computing where a similar decoupling has occurred. There, software has been, and continues to be, developed to enable infrastructure to be deployed dynamically adding and removing virtual servers. This has resulted in enterprises adopting these tools for their own internal infrastructure, commonly called a private cloud. Of course, here the meaning of infrastructure provider refers to the infrastructure being maintained by one business unit (the private cloud IT staff) and the customer refers to the various business units which are independently running services on the infrastructure. ShadowNet [6] is one such infrastructure being deployed at AT&T that is being used as a testbed for testing new services. It has an ultimate goal of being used for deploying new services, which, once it reaches that goal, it will essentially be a private cloud (but with the IaaS model - so it has some benefits, but not all). Outsourced IT Many small companies cannot afford the dedicated staff needed to manage their routers, yet need routers as normal course of business. This will become increasingly true as routers support more than just connectivity (as exemplified by the many innovative applications showcased in Cisco s application development competition, many of which were geared towards enterprises). Using the router platform, the ISP or 3rd party can simply provide a configuration of a router and automated tools map that to the customer s physical infrastructure. 6 Conclusions Like the shift that has taken place in computing towards hosted and shared architecture, we argue that the same will take place for networking, both in terms of business models and in terms of management of a private network. This will enable innovation, simplify management, and increase customer choice. We presented the router platform model where each customer is able to configure a single router (defined to include in-network general purpose processing). This single router is then automatically combined with other customers routers and mapped onto the distributed resources available. Moving to a new model will certainly be a challenge. We hope that the benefits of doing so will make it a reality. References [1] D. Taylor and J. Turner, Diversifying the Internet, in IEEE GLOBECOM, November [2] N. Feamster, L. Gao, and J. Rexford, How to lease the Internet in your spare time, in ACM Computer Communication Review, January [3] G. Schaffrath, C. Werle, P. Papadimitriou, A. Feldmann, R. Bless, A. Greenhalgh, M. Kind, O. Maennel, and L. Mathy, Network Virtualization Architecture: Proposal and Initial Prototype, in Workshop on Virtualized Infastructure Systems and Architectures (VISA), August [4] L. Peterson, S. Shenker, and J. Turner, Overcoming the Internet Impasse Through Virtualization, in Proceedings of the 3rd ACM Workshop on Hot Topics in Networks (HotNets-III), November [5] The Global Environment for Network Innovations (GENI), February [6] X. Chen, Z. M. Mao, and J. V. der Merwe, ShadowNet: A Platform for Rapid and Safe Network Evolution, in USENIX Annual Technical Conference, June [7] Y. Wang, E. Keller, B. Biskeborn, J. van der Merwe, and J. Rexford, Virtual Routers on the Move: Live Router Migration as a Network-Management Primitive, in ACM SIGCOMM, August [8] Y. Wang, M. Schapira, and J. Rexford, Neighbor-specific BGP: more flexible routing policies while improving global stability, in ACM SIGMETRICS, June [9] K. Lakshminarayanan, I. Stoica, and S. Shenker, Routing as a Service, Tech. Rep. UCB/CSD , UC Berkeley, January [10] A. Edwards, A. Fischer, and A. Lain, Diverter: A New Approach to Networking Within Virtualized Infrastructures, in Workshop on Research on Enterprise Networking (WREN), August [11] T. Wood, A. Gerber, K. Ramakrishnan, P. Shenoy, and J. V. der Merwe, The Case for Enterprise-Ready Virtual Private Clouds, in Workshop on Hot Topics in Cloud Computing (HotCloud), June [12] E. Keller, J. Rexford, and J. V. der Merwe, Seamless BGP Migration With Router Grafting, in Proc. Networked Systems Design and Implementation, April To appear. [13] Linux Containers - Network Namespaces. sourceforge.net/network.php,
Tag Switching. Background. Tag-Switching Architecture. Forwarding Component CHAPTER
CHAPTER 23 Tag Switching Background Rapid changes in the type (and quantity) of traffic handled by the Internet and the explosion in the number of Internet users is putting an unprecedented strain on the
More informationIntroduction to Segment Routing
Segment Routing (SR) is a flexible, scalable way of doing source routing. Overview of Segment Routing, page 1 How Segment Routing Works, page 2 Examples for Segment Routing, page 3 Benefits of Segment
More informationThe Case for Separating Routing from Routers
The Case for Separating Routing from Routers Nick Feamster, Hari Balakrishnan M.I.T. Computer Science and Artificial Intelligence Laboratory Jennifer Rexford, Aman Shaikh, Kobus van der Merwe AT&T Labs
More informationQOS Quality Of Service
QOS Quality Of Service Michael Schär Seminar in Distributed Computing Outline Definition QOS Attempts and problems in the past (2 Papers) A possible solution for the future: Overlay networks (2 Papers)
More informationDynamics of Hot-Potato Routing in IP Networks
Dynamics of Hot-Potato Routing in IP Networks Jennifer Rexford AT&T Labs Research http://www.research.att.com/~jrex Joint work with Renata Teixeira (UCSD), Aman Shaikh (AT&T), and Timothy Griffin (Intel)
More informationAdvanced Computer Networks
Advanced Computer Networks More on BGP Jianping Pan Summer 2007 7/4/07 csc485b/586b/seng480b 1 Review: BGP Border Gateway Protocol path vector routing prefix: AS-path policy-based routing import/export
More informationCisco 4000 Series Integrated Services Routers: Architecture for Branch-Office Agility
White Paper Cisco 4000 Series Integrated Services Routers: Architecture for Branch-Office Agility The Cisco 4000 Series Integrated Services Routers (ISRs) are designed for distributed organizations with
More informationMPLS in the DCN. Introduction CHAPTER
CHAPTER 5 First Published: January 3, 2008 Last Updated: January 3, 2008 Finding Support Information for Platforms and Cisco IOS and Catalyst OS Software Images Use Cisco Feature Navigator to find information
More informationNetworking for a smarter data center: Getting it right
IBM Global Technology Services October 2011 Networking for a smarter data center: Getting it right Planning the network needed for a dynamic infrastructure 2 Networking for a smarter data center: Getting
More informationAn Introduction to Overlay Networks PlanetLab: A Virtual Overlay Network Testbed
An Introduction to Overlay Networks PlanetLab: A Virtual Overlay Network Testbed Suhas Mathur suhas@winlab.rutgers.edu Communication Networks II Spring 2005 Talk Outline Introduction: The future internet
More informationThe Top Five Reasons to Deploy Software-Defined Networks and Network Functions Virtualization
The Top Five Reasons to Deploy Software-Defined Networks and Network Functions Virtualization May 2014 Prepared by: Zeus Kerravala The Top Five Reasons to Deploy Software-Defined Networks and Network Functions
More informationNetworking for a dynamic infrastructure: getting it right.
IBM Global Technology Services Networking for a dynamic infrastructure: getting it right. A guide for realizing the full potential of virtualization June 2009 Executive summary June 2009 Networking for
More informationCisco Webex Cloud Connected Audio
White Paper Cisco Webex Cloud Connected Audio Take full advantage of your existing IP telephony infrastructure to help enable a Webex integrated conferencing experience Introduction Cisco Webex Cloud Connected
More informationNetwork Implications of Cloud Computing Presentation to Internet2 Meeting November 4, 2010
Network Implications of Cloud Computing Presentation to Internet2 Meeting November 4, 2010 Lou Topfl Director, New Technology Product Development Engineering AT&T Agenda What is the Cloud? Types of Cloud
More informationMPLS VPN--Inter-AS Option AB
The feature combines the best functionality of an Inter-AS Option (10) A and Inter-AS Option (10) B network to allow a Multiprotocol Label Switching (MPLS) Virtual Private Network (VPN) service provider
More informationTELE 301 Network Management
TELE 301 Network Management Lecture 24: Exterior Routing and BGP Haibo Zhang Computer Science, University of Otago TELE301 Lecture 16: Remote Terminal Services 1 Today s Focus How routing between different
More informationIPv6 Switching: Provider Edge Router over MPLS
Multiprotocol Label Switching (MPLS) is deployed by many service providers in their IPv4 networks. Service providers want to introduce IPv6 services to their customers, but changes to their existing IPv4
More informationConfiguring MPLS L3VPN
Contents Configuring MPLS L3VPN 1 MPLS L3VPN overview 1 Introduction to MPLS L3VPN 1 MPLS L3VPN concepts 2 MPLS L3VPN packet forwarding 5 MPLS L3VPN networking schemes 5 MPLS L3VPN routing information
More informationNEXT GENERATION CMTS CHARACTERISTICS INCLUDING IP MULTICAST. Doug Jones YAS Broadband Ventures
NEXT GENERATION CMTS CHARACTERISTICS INCLUDING IP MULTICAST Doug Jones YAS Broadband Ventures Abstract The cable industry is currently faced with upgrading from DOCSIS v1.0 CMTS equipment to DOCSIS v1.1
More informationCentralization of Network using Openflow Protocol
Indian Journal of Science and Technology, Vol 8(S2), 165 170, January 2015 ISSN (Print) : 0974-6846 ISSN (Online) : 0974-5645 DOI : 10.17485/ijst/2015/v8iS2/61217 Centralization of Network using Openflow
More informationConfiguring multicast VPN
Contents Configuring multicast VPN 1 Multicast VPN overview 1 Multicast VPN overview 1 MD-VPN overview 3 Protocols and standards 6 How MD-VPN works 6 Share-MDT establishment 6 Share-MDT-based delivery
More informationInterdomain Routing. Networked Systems (H) Lecture 11
Interdomain Routing Networked Systems (H) Lecture 11 Lecture Outline Interdomain routing Autonomous systems and the Internet AS-level topology BGP and Internet routing 2 Interdomain Unicast Routing Tier-1
More informationSoftware Defined Networking: A New Networking Approach
Software Defined Networking: A New Networking Approach Gunjan Katwal #, Ayushi Nainwal * # Computer Science, AP Goyal University Shimla 1 gun363@gmail.com 2 AyushiNainwal@gmail.com ABSTRACT Software defined
More informationMPLS L3VPN. The MPLS L3VPN model consists of three kinds of devices: PE CE Site 2. Figure 1 Network diagram for MPLS L3VPN model
is a kind of PE-based L3VPN technology for service provider VPN solutions. It uses BGP to advertise VPN routes and uses to forward VPN packets on service provider backbones. provides flexible networking
More informationRouters / external connectivity (HSRP) Web farm, mail servers
Routers / external connectivity (HSRP) hubs/switches Office network!#"%$'&)(+*-,/.10#23*-&4$5!6$5!7&)(6879:(;&
More informationCS4700/CS5700 Fundamentals of Computer Networks
CS4700/CS5700 Fundamentals of Computer Networks Lecture 22: Overlay networks Slides used with permissions from Edward W. Knightly, T. S. Eugene Ng, Ion Stoica, Hui Zhang Alan Mislove amislove at ccs.neu.edu
More informationIPv6 Switching: Provider Edge Router over MPLS
Multiprotocol Label Switching (MPLS) is deployed by many service providers in their IPv4 networks. Service providers want to introduce IPv6 services to their customers, but changes to their existing IPv4
More informationCisco Nexus 9200 Switch Datasheet
Cisco Nexus 9200 Switch Datasheet CONTENT Content... 1 Overview... 2 Appearance... 2 Key Features and Benefits... 3 NX-OS Software... 4 Nexus 9200 Compare models... 6 Specification of nexus 9200 series
More informationMPLS VPN Inter-AS Option AB
First Published: December 17, 2007 Last Updated: September 21, 2011 The feature combines the best functionality of an Inter-AS Option (10) A and Inter-AS Option (10) B network to allow a Multiprotocol
More informationMASERGY S MANAGED SD-WAN
MASERGY S MANAGED New Performance Options for Hybrid Networks Business Challenges WAN Ecosystem Features and Benefits Use Cases INTRODUCTION Organizations are leveraging technology to transform the way
More informationCompetitive and Deterministic Embeddings of Virtual Networks
Competitive and Deterministic Embeddings of Virtual Networks Guy Even (Tel Aviv Uni) Moti Medina (Tel Aviv Uni) Gregor Schaffrath (T-Labs Berlin) Stefan Schmid (T-Labs Berlin) The Virtual Network Embedding
More informationIntroduction to Cisco ASR 9000 Series Network Virtualization Technology
White Paper Introduction to Cisco ASR 9000 Series Network Virtualization Technology What You Will Learn Service providers worldwide face high customer expectations along with growing demand for network
More informationAdvanced Multihoming. BGP Traffic Engineering
Advanced Multihoming BGP Traffic Engineering 1 Service Provider Multihoming Previous examples dealt with loadsharing inbound traffic Of primary concern at Internet edge What about outbound traffic? Transit
More informationMulti Topology Routing Truman Boyes
Multi Topology Routing Truman Boyes truman@juniper.net Copyright 2008 Juniper Networks, Inc. 1 Traffic Engineering Choices Today: IGP Metric Costing RSVP TE end to end Policy based routing EROs, Offline
More informationVXLAN Design with Cisco Nexus 9300 Platform Switches
Guide VXLAN Design with Cisco Nexus 9300 Platform Switches Guide October 2014 2014 Cisco and/or its affiliates. All rights reserved. This document is Cisco Public Information. Page 1 of 39 Contents What
More informationSoftware-Defined Networking (SDN) Overview
Reti di Telecomunicazione a.y. 2015-2016 Software-Defined Networking (SDN) Overview Ing. Luca Davoli Ph.D. Student Network Security (NetSec) Laboratory davoli@ce.unipr.it Luca Davoli davoli@ce.unipr.it
More informationInter-Domain Routing: BGP
Inter-Domain Routing: BGP Brad Karp UCL Computer Science (drawn mostly from lecture notes by Hari Balakrishnan and Nick Feamster, MIT) CS 3035/GZ01 4 th December 2014 Outline Context: Inter-Domain Routing
More informationUnifying the Distributed Enterprise with MPLS Mesh
Unifying the Distributed Enterprise with MPLS Mesh Technical Whitepaper January 2015 Copyright 2015 AireSpring Introduction Today s modern enterprises employ IT technologies that deliver higher value,
More informationGuide to SDN, SD-WAN, NFV, and VNF
Evalu- ation Guide Technical Brief SD-WAN, NFV, and With so many acronyms and each one building on another, it can be confusing about how they work together. This guide describes the difference. 006180206
More informationResource analysis of network virtualization through user and network
Resource analysis of network virtualization through user and network P.N.V.VAMSI LALA, Cloud Computing, Master of Technology, SRM University, Potheri. Mr.k.Venkatesh, Asst.Professor (Sr.G), Information
More informationService Provider Multihoming
Service Provider Multihoming ISP Workshops These materials are licensed under the Creative Commons Attribution-NonCommercial 4.0 International license (http://creativecommons.org/licenses/by-nc/4.0/) Last
More informationETSI FUTURE Network SDN and NFV for Carriers MP Odini HP CMS CT Office April 2013
ETSI FUTURE Network SDN and NFV for Carriers MP Odini HP CMS CT Office April 2013 Challenges and Opportunities Reduce Capex/Opex Challenges Innovation at devices and OTT side Number of devices explode
More informationSoftware-Defined Networking. Daphné Tuncer Department of Computing Imperial College London (UK)
Software-Defined Networking Daphné Tuncer Department of Computing Imperial College London (UK) dtuncer@ic.ac.uk 25/10/2018 Agenda Part I: Principles of Software-Defined Networking (SDN) 1. Why a lecture
More informationTechnical Document. What You Need to Know About Ethernet Audio
Technical Document What You Need to Know About Ethernet Audio Overview Designing and implementing an IP-Audio Network can be a daunting task. The purpose of this paper is to help make some of these decisions
More informationMPLS VPN C H A P T E R S U P P L E M E N T. BGP Advertising IPv4 Prefixes with a Label
7 C H A P T E R S U P P L E M E N T This online supplement of Chapter 7 focuses on two important developments. The first one is Inter-Autonomous. Inter-Autonomous is a concept whereby two service provider
More informationA configuration-only approach to shrinking FIBs. Prof Paul Francis (Cornell)
A configuration-only approach to shrinking FIBs Prof Paul Francis (Cornell) 1 Virtual Aggregation An approach to shrinking FIBs (and RIBs) In routers, not in route reflectors Works with legacy routers
More informationVXLAN Overview: Cisco Nexus 9000 Series Switches
White Paper VXLAN Overview: Cisco Nexus 9000 Series Switches What You Will Learn Traditional network segmentation has been provided by VLANs that are standardized under the IEEE 802.1Q group. VLANs provide
More informationInterdomain Routing Reading: Sections K&R EE122: Intro to Communication Networks Fall 2007 (WF 4:00-5:30 in Cory 277)
Interdomain Routing Reading: Sections K&R 4.6.3 EE122: Intro to Communication Networks Fall 2007 (WF 4:00-5:30 in Cory 277) Guest Lecture by Brighten Godfrey Instructor: Vern Paxson TAs: Lisa Fowler, Daniel
More informationIPv6 Module 4 OSPF to IS-IS for IPv6
IPv6 Module 4 OSPF to IS-IS for IPv6 Objective: To migrate the OSPF version of Module 1 (running IPv4) to using IS-IS as part of an IPv6 migration strategy. OSPF will be completely removed once the migration
More informationMiddleboxes. CSU CS557 - Fall 2017 Instructor: Lorenzo De Carli
Middleboxes CSU CS557 - Fall 2017 Instructor: Lorenzo De Carli What is a middlebox? Middlebox : networking parlance for any network device which performs analysis and/or transformation of application-level
More informationIP Addressing & Interdomain Routing. Next Topic
IP Addressing & Interdomain Routing Next Topic IP Addressing Hierarchy (prefixes, class A, B, C, subnets) Interdomain routing Application Presentation Session Transport Network Data Link Physical Scalability
More informationSecurizarea Calculatoarelor și a Rețelelor 32. Tehnologia MPLS VPN
Platformă de e-learning și curriculă e-content pentru învățământul superior tehnic Securizarea Calculatoarelor și a Rețelelor 32. Tehnologia MPLS VPN MPLS VPN 5-ian-2010 What this lecture is about: IP
More informationSimplifying the Branch Network
Simplifying the Branch Network By: Lee Doyle, Principal Analyst at Doyle Research Sponsored by Aruba, a Hewlett Packard Enterprise company Executive Summary A majority of IT organizations are experiencing
More informationService Provider Multihoming
Service Provider Multihoming ISP Workshops Last updated 18 September 2013 1 Service Provider Multihoming p Previous examples dealt with loadsharing inbound traffic n Of primary concern at Internet edge
More informationCS 4226: Internet Architecture
Software Defined Networking Richard T. B. Ma School of Computing National University of Singapore Material from: Scott Shenker (UC Berkeley), Nick McKeown (Stanford), Jennifer Rexford (Princeton) CS 4226:
More informationBGP in the Internet Best Current Practices
BGP in the Internet Best Current Practices 1 Recommended IOS Releases Which IOS?? 2 Which IOS? IOS is a feature rich and highly complex router control system ISPs should choose the IOS variant which is
More informationSoftware Defined Networks and OpenFlow. Courtesy of: AT&T Tech Talks.
MOBILE COMMUNICATION AND INTERNET TECHNOLOGIES Software Defined Networks and Courtesy of: AT&T Tech Talks http://web.uettaxila.edu.pk/cms/2017/spr2017/temcitms/ MODULE OVERVIEW Motivation behind Software
More informationSDN Use-Cases. internet exchange, home networks. TELE4642: Week8. Materials from Prof. Nick Feamster is gratefully acknowledged
SDN Use-Cases internet exchange, home networks TELE4642: Week8 Materials from Prof. Nick Feamster is gratefully acknowledged Overview n SDX: A Software-Defined Internet Exchange n SDN-enabled Home Networks
More informationOPEN COMPUTE PLATFORMS POWER SOFTWARE-DRIVEN PACKET FLOW VISIBILITY, PART 2 EXECUTIVE SUMMARY. Key Takeaways
OPEN COMPUTE PLATFORMS POWER SOFTWARE-DRIVEN PACKET FLOW VISIBILITY, PART 2 EXECUTIVE SUMMARY This is the second of two white papers that describe how the shift from monolithic, purpose-built, network
More informationInterdomain Routing Reading: Sections P&D 4.3.{3,4}
Interdomain Routing Reading: Sections P&D 4.3.{3,4} EE122: Intro to Communication Networks Fall 2006 (MW 4:00-5:30 in Donner 155) Vern Paxson TAs: Dilip Antony Joseph and Sukun Kim http://inst.eecs.berkeley.edu/~ee122/
More informationMPLS/Tag Switching. Background. Chapter Goals CHAPTER
28 CHAPTER Chapter Goals Understand the advantages of MPLS. Learn the components of an MPLS system. Compare and contrast MPLS and hop-by-hop routing. Describe the two methods of label distribution. Explain
More informationDeployments and Network Topologies
TECHNICAL GUIDE Deployments and Network Topologies A technical guide to deploying Family Zone School in different network topologies. Contents Introduction...........................................3 Transparent
More informationFinding a Needle in a Haystack: Pinpointing Significant BGP Routing Changes in an IP Network
Finding a Needle in a Haystack: Pinpointing Significant BGP Routing Changes in an IP Network Jian Wu (University of Michigan) Z. Morley Mao (University of Michigan) Jennifer Rexford (Princeton University)
More informationLecture 4: Intradomain Routing. CS 598: Advanced Internetworking Matthew Caesar February 1, 2011
Lecture 4: Intradomain Routing CS 598: Advanced Internetworking Matthew Caesar February 1, 011 1 Robert. How can routers find paths? Robert s local DNS server 10.1.8.7 A 10.1.0.0/16 10.1.0.1 Routing Table
More informationImprove Multiplayer Online Games
Games: Leveraging SDN Networks to Improve Multiplayer Online Games Shufeng Huang Laboratory for Advanced Networking University of Kentucky Lexington, KY 40506 Email: shufeng@netlab.uky.edu James Griffioen
More informationCloud Services. Infrastructure-as-a-Service
Cloud Services Infrastructure-as-a-Service Accelerate your IT and business transformation with our networkcentric, highly secure private and public cloud services - all backed-up by a 99.999% availability
More informationCS4450. Computer Networks: Architecture and Protocols. Lecture 15 BGP. Spring 2018 Rachit Agarwal
CS4450 Computer Networks: Architecture and Protocols Lecture 15 BGP Spring 2018 Rachit Agarwal Autonomous System (AS) or Domain Region of a network under a single administrative entity Border Routers Interior
More informationCompetitive Public Switched Telephone Network (PSTN) Wide- Area Network (WAN) Access Using Signaling System 7 (SS7)
Competitive Public Switched Telephone Network (PSTN) Wide- Area Network (WAN) Access Using Signaling System 7 (SS7) Definition Using conventional Internet access equipment, service providers may access
More informationData Center Network Topologies II
Data Center Network Topologies II Hakim Weatherspoon Associate Professor, Dept of Computer cience C 5413: High Performance ystems and Networking April 10, 2017 March 31, 2017 Agenda for semester Project
More informationSaaS Providers. ThousandEyes for. Summary
USE CASE ThousandEyes for SaaS Providers Summary With Software-as-a-Service (SaaS) applications rapidly replacing onpremise solutions, the onus of ensuring a great user experience for these applications
More informationData Center Virtualization: VirtualWire
Data Center Virtualization: VirtualWire Hakim Weatherspoon Assistant Professor, Dept of Computer Science CS 5413: High Performance Systems and Networking November 21, 2014 Slides from USENIX Workshop on
More informationCisco Wide Area Application Services and Cisco Nexus Family Switches: Enable the Intelligent Data Center
Cisco Wide Area Application Services and Cisco Nexus Family Switches: Enable the Intelligent Data Center What You Will Learn IT departments are facing increasing pressure to accommodate numerous changing
More informationLocator ID Separation Protocol (LISP) Overview
Locator ID Separation Protocol (LISP) is a network architecture and protocol that implements the use of two namespaces instead of a single IP address: Endpoint identifiers (EIDs) assigned to end hosts.
More informationOverview. Information About Layer 3 Unicast Routing. Send document comments to CHAPTER
CHAPTER 1 This chapter introduces the basic concepts for Layer 3 unicast routing protocols in Cisco NX-OS. This chapter includes the following sections: Information About Layer 3 Unicast Routing, page
More informationFIGURE 3. Two-Level Internet Address Structure. FIGURE 4. Principle Classful IP Address Formats
Classful IP Addressing When IP was first standardized in September 1981, the specification required that each system attached to an IP-based Internet be assigned a unique, 32-bit Internet address value.
More informationThe Emerging Role of a CDN in Facilitating Secure Cloud Deployments
White Paper The Emerging Role of a CDN in Facilitating Secure Cloud Deployments Sponsored by: Fastly Robert Ayoub August 2017 IDC OPINION The ongoing adoption of cloud services and the desire for anytime,
More informationNever Drop a Call With TecInfo SIP Proxy White Paper
Innovative Solutions. Trusted Performance. Intelligently Engineered. Never Drop a Call With TecInfo SIP Proxy White Paper TecInfo SD-WAN product - PowerLink - enables real time traffic like VoIP, video
More informationVirtual Multi-homing: On the Feasibility of Combining Overlay Routing with BGP Routing
Virtual Multi-homing: On the Feasibility of Combining Overlay Routing with BGP Routing Zhi Li, Prasant Mohapatra, and Chen-Nee Chuah University of California, Davis, CA 95616, USA {lizhi, prasant}@cs.ucdavis.edu,
More informationA Scalable, Commodity Data Center Network Architecture
A Scalable, Commodity Data Center Network Architecture B Y M O H A M M A D A L - F A R E S A L E X A N D E R L O U K I S S A S A M I N V A H D A T P R E S E N T E D B Y N A N X I C H E N M A Y. 5, 2 0
More informationOverlay and P2P Networks. Introduction and unstructured networks. Prof. Sasu Tarkoma
Overlay and P2P Networks Introduction and unstructured networks Prof. Sasu Tarkoma 14.1.2013 Contents Overlay networks and intro to networking Unstructured networks Overlay Networks An overlay network
More informationInternet Routing Protocols Lecture 01 & 02
Internet Routing Protocols Lecture 01 & 02 Advanced Systems Topics Lent Term, 2010 Timothy G. Griffin Computer Lab Cambridge UK Internet Routing Outline Lecture 1 : Inter-domain routing architecture, the
More informationModeling the Routing of an ISP with C-BGP
Modeling the Routing of an ISP with C-BGP Bruno Quoitin bruno.quoitin@uclouvain.be IP Networking Lab (INL) Computer Science & Engineering Department Université catholique de Louvain, Belgium 2009 B. Quoitin
More informationInterdomain Routing. EE122 Fall 2011 Scott Shenker
Interdomain Routing EE122 Fall 2011 Scott Shenker http://inst.eecs.berkeley.edu/~ee122/ Materials with thanks to Jennifer Rexford, Ion Stoica, Vern Paxson and other colleagues at Princeton and UC Berkeley
More informationConfiguring MPLS L3VPN
Contents Configuring MPLS L3VPN 1 MPLS L3VPN overview 1 MPLS L3VPN concepts 2 MPLS L3VPN packet forwarding 4 MPLS L3VPN networking schemes 5 MPLS L3VPN routing information advertisement 8 Inter-AS VPN
More informationTALK THUNDER SOFTWARE FOR BARE METAL HIGH-PERFORMANCE SOFTWARE FOR THE MODERN DATA CENTER WITH A10 DATASHEET YOUR CHOICE OF HARDWARE
DATASHEET THUNDER SOFTWARE FOR BARE METAL YOUR CHOICE OF HARDWARE A10 Networks application networking and security solutions for bare metal raise the bar on performance with an industryleading software
More information15-744: Computer Networking. Middleboxes and NFV
15-744: Computer Networking Middleboxes and NFV Middleboxes and NFV Overview of NFV Challenge of middleboxes Middlebox consolidation Outsourcing middlebox functionality Readings: Network Functions Virtualization
More informationMPLS Networks: Design and Routing Functions
MPLS Networks: Design and Routing Functions Course Description This course provides an understanding of how MPLS works its advantages and limitations and how it can be deployed to provide effective services
More informationHow to Configure a Hybrid WAN in Parallel to An Existing Traditional Wan Infrastructure
Table of Contents Introducing Hybrid WAN The WAN Evolution What is a Traditional Private WAN? Example IP Addressing What is a Hybrid WAN? SD-WAN - Centrally Managed Hybrid WAN Hybrid WAN Example Hybrid
More informationA Better Approach to Leveraging an OpenStack Private Cloud. David Linthicum
A Better Approach to Leveraging an OpenStack Private Cloud David Linthicum A Better Approach to Leveraging an OpenStack Private Cloud 1 Executive Summary The latest bi-annual survey data of OpenStack users
More informationWhite. Paper. EMC Isilon Scale-out Without Compromise. July, 2012
White Paper EMC Isilon Scale-out Without Compromise By Terri McClure, Senior Analyst July, 2012 This ESG White Paper was commissioned by EMC and is distributed under license from ESG. 2012, The Enterprise
More informationBGP IN THE DATA CENTER
BGP IN THE DATA CENTER A PACKET DESIGN E-BOOK Contents Page 3 : BGP the Savior Page 4 : Traditional Data Center Architecture Traffic Flows Scalability Spanning Tree Protocol (STP) Page 6 : CLOS Architecture
More informationMPLS VPN Carrier Supporting Carrier Using LDP and an IGP
MPLS VPN Carrier Supporting Carrier Using LDP and an IGP Multiprotocol Label Switching (MPLS) Virtual Private Network (VPN) Carrier Supporting Carrier (CSC) enables one MPLS VPN-based service provider
More informationA Cloud WHERE PHYSICAL ARE TOGETHER AT LAST
A Cloud WHERE PHYSICAL AND VIRTUAL STORAGE ARE TOGETHER AT LAST Not all Cloud solutions are the same so how do you know which one is right for your business now and in the future? NTT Communications ICT
More informationIPv6 Module 6x ibgp and Basic ebgp
IPv6 Module 6x ibgp and Basic ebgp Objective: Using IPv6, simulate four different interconnected ISP backbones using a combination of IS-IS, internal BGP, and external BGP. Topology : Figure 1 BGP AS Numbers
More informationOvercoming the Internet Impasse through Virtualization Thomas Anderson, Larry Peterson, Scott Shenker, Jonathan Turner. 원종호 (INC lab) Sep 25, 2006
Overcoming the Internet Impasse through Virtualization Thomas Anderson, Larry Peterson, Scott Shenker, Jonathan Turner 원종호 (INC lab) Sep 25, 2006 Outline Introduction Three requirements Virtualization
More informationPathlet Routing. P. Brighten Godfrey, Igor Ganichev, Scott Shenker, and Ion Stoica SIGCOMM (maurizio patrignani)
Pathlet Routing P. Brighten Godfrey, Igor Ganichev, Scott Shenker, and Ion Stoica SIGCOMM 2009 (maurizio patrignani) Reti di Calcolatori di Nuova Generazione http://www.dia.uniroma3.it/~rimondin/courses/rcng1011/
More informationCisco Data Center Network Manager 5.1
Cisco Data Center Network Manager 5.1 Product Overview Modern data centers are becoming increasingly large and complex. New technology architectures such as cloud computing and virtualization are adding
More informationTransform your network and your customer experience. Introducing SD-WAN Concierge
Transform your network and your customer experience Introducing SD-WAN Concierge Optimize your application performance, lower your total cost of ownership and simplify your network management. 2X Bandwith
More informationManaging Costs in Growing Networks with SEGway STPs and Point Code Emulation
Managing Costs in Growing Networks with SEGway STPs and Point Code Emulation Deb Brunner-Walker, Performance Technologies Technology White Paper www.pt.com Introduction Voice networks are evolving from
More informationMultihoming with BGP and NAT
Eliminating ISP as a single point of failure www.noction.com Table of Contents Introduction 1. R-NAT Configuration 1.1 NAT Configuration 5. ISPs Routers Configuration 3 15 7 7 5.1 ISP-A Configuration 5.2
More information